[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » user » zeman » treebanks

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
user:zeman:treebanks [2011/11/03 11:16]
zeman Parsing results. 		user:zeman:treebanks [2014/07/17 17:43] (current)
zeman Croatian.
Line 1: Line 1:
 ====== Treebanks for Various Languages ====== ====== Treebanks for Various Languages ======
  
-===== Arabic (ar) =====+http://ufal.mff.cuni.cz/hamledt/ nebo [[hamledt|HamleDT ve Wiki]]
  
-Prague Arabic Dependency Treebank (PADT)+  * [[user:zeman:treebanks:grc|Ancient Greek (grc)]] 
 +  * [[user:zeman:treebanks:ar|Arabic (ar)]] 
 +  * [[user:zeman:treebanks:eu|Basque (eu)]] 
 +  * [[user:zeman:treebanks:bn|Bengali (bn)]] 
 +  * [[user:zeman:treebanks:bg|Bulgarian (bg)]] 
 +  * [[user:zeman:treebanks:ca|Catalan (ca)]] 
 +  * [[user:zeman:treebanks:hr|Croatian (hr)]] 
 +  * [[user:zeman:treebanks:cs|Czech (cs)]] 
 +  * [[user:zeman:treebanks:da|Danish (da)]] 
 +  * [[user:zeman:treebanks:nl|Dutch (nl)]] 
 +  * [[user:zeman:treebanks:en|English (en)]] 
 +  * [[user:zeman:treebanks:et|Estonian (et)]] 
 +  * [[user:zeman:treebanks:fi|Finnish (fi)]] 
 +  * [[user:zeman:treebanks:de|German (de)]] 
 +  * [[user:zeman:treebanks:el|Greek (el)]] 
 +  * [[user:zeman:treebanks:hi|Hindi (hi)]] 
 +  * [[user:zeman:treebanks:hu|Hungarian (hu)]] 
 +  * [[user:zeman:treebanks:it|Italian (it)]] 
 +  * [[user:zeman:treebanks:ja|Japanese (ja)]] 
 +  * [[user:zeman:treebanks:la|Latin (la)]] 
 +  * [[user:zeman:treebanks:fa|Persian (fa)]] 
 +  * [[user:zeman:treebanks:pt|Portuguese (pt)]] 
 +  * [[user:zeman:treebanks:ro|Romanian (ro)]] 
 +  * [[user:zeman:treebanks:ru|Russian (ru)]] 
 +  * [[user:zeman:treebanks:sk|Slovak (sk)]] 
 +  * [[user:zeman:treebanks:sl|Slovene (sl)]] 
 +  * [[user:zeman:treebanks:es|Spanish (es)]] 
 +  * [[user:zeman:treebanks:sv|Swedish (sv)]] 
 +  * [[user:zeman:treebanks:ta|Tamil (ta)]] 
 +  * [[user:zeman:treebanks:te|Telugu (te)]] 
 +  * [[user:zeman:treebanks:tr|Turkish (tr)]]
  
-==== Versions ====+===== To Process =====
  
-  * Original PADT 1.0 as distributed by the LDC +Ahoj, 
-  * CoNLL 2006 +stáhl jsem nový španělský závislostní korpus IULA (větší než AnCora) 
-  * CoNLL 2007+/net/projects/tectomt_shared/data/resources/treebanks/es
  
-The CoNLL 2007 version reportedly improves over CoNLL 2006 in quality of morphological annotation. Both CoNLL versions miss important parts of the original PADT annotation (see below).+License:  CC BY 3.0 (Unported) 
 +Web:      http://www.iula.upf.edu/recurs01_tbk_uk.htm 
 +Doc:      http://www.iula.upf.edu/recurs01_conll_uk.htm 
 +Download: http://repositori.upf.edu/handle/10230/20048 
 +Parsing:  http://www.taln.upf.edu/system/files/biblio_files/ijcnlp_final_padro_et_al_2013.pdf 
 +          state-of-the-art LAS score is 94.7 using Mate-C 
 +sentences  42,000 
 +tokens    590,000
  
-==== Obtaining and License ==== +The sentences have been choosed from the IULA LSP corpus, automatically annotated with POS information and manually annotated with syntactical information using the DELPH-IN environment. The resulting syntactic analysis is automatically converted to dependencies and delivered using the CONLL format.
- +
-The original PADT 1.0 is distributed by the LDC under the catalogue number [[http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2004T23|LDC2004T23]]. It is free for LDC members 2004, price for non-members is unknown (contact LDC). The [[http://ufal.mff.cuni.cz/corp-lic/padt10-ord.html|license]] in short: +
- +
-  * non-commercial research usage +
-  * no redistribution +
-  * cite [[http://ufal.mff.cuni.cz/padt/PADT_1.0/#refs|one publication]] in publications; also include: "The PADT 1.0 has been developed by the Institute of Formal and Applied Linguistics and the Center for Computational Linguistics (see http://ufal.mff.cuni.cz/)." +
- +
-The CoNLL 2006 and 2007 versions are obtainable upon request under similar license terms. Their publication in the LDC together with the other CoNLL treebanks is being prepared. +
- +
-PADT was created by members of the [[http://ufal.mff.cuni.cz/|Institute of Formal and Applied Linguistics]] (Ústav formální a aplikované lingvistiky, ÚFAL), Faculty of Mathematics and Physics (Matematicko-fyzikální fakulta), Charles University in Prague (Univerzita Karlova v Praze), Malostranské náměstí 25, Praha, CZ-11800, Czechia. +
- +
-==== Domain ==== +
- +
-Newswire text from press agencies (Agence France Presse, Ummah, Al Hayat, An Nahar, Xinhua 2001-2003). +
- +
-==== Size ==== +
- +
-According to their website, the original PADT 1.0 contains 113,500 tokens annotated analytically. The CoNLL 2006 version contains 59752 tokens in 1606 sentences, yielding 37.21 tokens per sentence on average (CoNLL 2006 data split: 54379 tokens / 1460 sentences training, 5373 tokens / 146 sentences test). The CoNLL 2007 version contains 116,793 tokens in 3043 sentences, yielding 38.38 tokens per sentence on average (CoNLL 2007 data split: 111,669 tokens / 2912 sentences training, 5124 tokens / 131 sentences test). +
- +
-==== References ==== +
- +
-  * Website +
-    * http://ufal.mff.cuni.cz/padt/ +
-  * Data +
-    * Jan Hajič, Otakar Smrž, Petr Zemánek, Petr Pajas, Jan Šnaidauf, Emanuel Beška, Jakub Kráčmar, Kamila Hassanová: //Prague Arabic Dependency Treebank 1.0// (LDC2004T23). Linguistic Data Consortium, Philadelphia, USA, 2004. ISBN 1-58563-319-4. +
-  * Principal publications +
-    * Jan Hajič, Otakar Smrž, Petr Zemánek, Jan Šnaidauf, Emanuel Beška: [[http://ufal.mff.cuni.cz/padt/PADT_1.0/docs/papers/2004-nemlar-padt.pdf|Prague Arabic Dependency Treebank: Development in Data and Tools]]. In: //Proceedings of the NEMLAR International Conference on Arabic Language Resources and Tools//, pages 110-117, Cairo, Egypt, 2004. +
- +
-==== Inside ==== +
- +
-The original PADT 1.0 is distributed in the [[:format-fs|FS format]]. The CoNLL versions are distributed in the [[:format-conll|CoNLL-X format]]. The original PADT contains more information than the CoNLL version. There is morphological annotation (tags and lemmas) both manual and by a tagger (while only manual is in CoNLL data), glosses etc. However, the most important piece of information that got lost during the conversion to CoNLL is the FS attribute called ''parallel''. It distinguishes conjuncts from shared modifiers of coordination and thus the syntactic structure is incomplete without it. +
- +
-Word forms and lemmas are vocalized, i.e. they contain diacritics for short vowels as well as consonant gemination and a few other things. The CoNLL 2006 version includes [[http://www.qamus.org/transliteration.htm|Buckwalter transliteration]] of the Arabic script (in the same column as Arabic, attached to the Arabic form/lemma with an underscore character). +
- +
-Note that tokenization of Arabic typically includes splitting original words (inserting spaces between letters), not just separating punctuation from words. Example: وبالفالوجة = wabiālfālūjah = wa/CONJ + bi/PREP + AlfAlwjp/NOUN_PROP = and in al-Falujah. In PADT, conjunctions and prepositions are separate tokens and nodes. +
- +
-The original PADT 1.0 uses 10-character positional morphological tags whose documentation is hard to find. The CoNLL 2006 version converts the tags to the three CoNLL columns, CPOS, POS and FEAT, most of the information being encoded as pipe-separated attribute-value assignments in FEAT. There //should// be a 1-1 mapping between the PADT positional tags and the CoNLL 2006 annotation. The CoNLL 2007 version uses a tag conversion different from CoNLL 2006. Both CoNLL distributions contain a README file with a brief description of the parts of speech and features. Use [[http://quest.ms.mff.cuni.cz/cgi-bin/interset/index.pl?tagset=ar::conll|DZ Interset]] to inspect the two CoNLL tagsets. Also look at the [[http://quest.ms.mff.cuni.cz/cgi-bin/elixir/index.fcgi|Elixir FM online interface]] for a later development of the morphological analyzer created along with PADT. +
- +
-The guidelines for syntactic annotation are documented in the [[http://ufal.mff.cuni.cz/padt/PADT_1.0/docs/guides/PADT_Analytical.pdf|PADT annotation manual]] (only peculiarities of Arabic are documented, otherwise it is referenced to the annotation manual for the Czech treebank). The list and brief description of syntactic tags (dependency relation labels) can be found in [[http://ufal.mff.cuni.cz/padt/PADT_1.0/docs/papers/2002-flm-padt.pdf|(Smrž, Šnaidauf and Zemánek, 2002)]]. +
- +
-==== Sample ==== +
- +
-The first two sentences of the CoNLL 2006 training data: +
- +
-| 1 | غِيابُ_giyAbu | غِياب_giyAb | N | N | case=1<nowiki>|</nowiki>def=R | 0 | ExD | _ | _ | +
-| 2 | فُؤاد_fu&Ad | فُؤاد_fu&Ad | Z | Z | _ | 3 | Atr | _ | _ | +
-| 3 | كَنْعان_kanoEAn | كَنْعان_kanoEAn | Z | Z | _ | 1 | Atr | _ | _ | +
-| |||||||||| +
-| 1 | فُؤاد_fu&Ad | فُؤاد_fu&Ad | Z | Z | _ | 2 | Atr | _ | _ | +
-| 2 | كَنْعان_kanoEAn | كَنْعان_kanoEAn | Z | Z | _ | 9 | Sb | _ | _ | +
-| 3 | ،_, | ،_, | G | G | _ | 2 | AuxG | _ | _ | +
-| 4 | رائِد_rA}id | رائِد_rA}id | N | N | _ | 2 | Atr | _ | _ | +
-| 5 | القِصَّة_AlqiS~ap | قِصَّة_qiS~ap | N | N | gen=F<nowiki>|</nowiki>num=S<nowiki>|</nowiki>def=D | 4 | Atr | _ | _ | +
-| 6 | القَصِيرَةِ_AlqaSiyrapi | قَصِير_qaSiyr | A | A | gen=F<nowiki>|</nowiki>num=S<nowiki>|</nowiki>case=2<nowiki>|</nowiki>def=D | 5 | Atr | _ | _ | +
-| 7 | فِي_fiy | فِي_fiy | P | P | _ | 4 | AuxP | _ | _ | +
-| 8 | لُبْنانِ_lubonAni | لُبْنان_lubonAn | Z | Z | case=2<nowiki>|</nowiki>def=R | 7 | Atr | _ | _ | +
-| 9 | رَحَلَ_raHala | رَحَل-َ_raHal-a | V | VP | pers=3<nowiki>|</nowiki>gen=M<nowiki>|</nowiki>num=S | 0 | Pred | _ | _ | +
-| 10 | مَساءَ_masA'a | مَساء_masA' | D | D | _ | 9 | Adv | _ | _ | +
-| 11 | أَمْسِ_>amosi | أَمْسِ_>amosi | D | D | _ | 10 | Atr | _ | _ | +
-| 12 | عَن_Ean | عَن_Ean | P | P | _ | 9 | AuxP | _ | _ | +
-| 13 | 81_81 | 81_81 | Q | Q | _ | 12 | Adv | _ | _ | +
-| 14 | عاماً_EAmAF | عام_EAm | N | N | gen=M<nowiki>|</nowiki>num=S<nowiki>|</nowiki>case=4<nowiki>|</nowiki>def=I | 13 | Atr | _ | _ | +
-| 15 | ._. | ._. | G | G | _ | 0 | AuxK | _ | _ | +
- +
-The first sentence of the CoNLL 2006 test data: +
- +
-| 1 | اِتِّفاقٌ_Ait~ifAqN | اِتِّفاق_Ait~ifAq | N | N | case=1<nowiki>|</nowiki>def=I | 0 | ExD | _ | _ | +
-| 2 | بَيْنَ_bayona | بَيْنَ_bayona | P | P | _ | 1 | AuxP | _ | _ | +
-| 3 | لُبْنانِ_lubonAni | لُبْنان_lubonAn | Z | Z | case=2<nowiki>|</nowiki>def=R | 4 | Atr | _ | _ | +
-| 4 | وَ_wa | وَ_wa | C | C | _ | 2 | Coord | _ | _ | +
-| 5 | سُورِيَّةٍ_suwriy~apK | سُورِيا_suwriyA | Z | Z | gen=F<nowiki>|</nowiki>num=S<nowiki>|</nowiki>case=2<nowiki>|</nowiki>def=I | 4 | Atr | _ | _ | +
-| 6 | عَلَى_EalaY | عَلَى_EalaY | P | P | _ | 1 | AuxP | _ | _ | +
-| 7 | رَفْعِ_rafoEi | رَفْع_rafoE | N | N | case=2<nowiki>|</nowiki>def=R | 6 | Atr | _ | _ | +
-| 8 | مُسْتَوَى_musotawaY | مُسْتَوَى_musotawaY | N | N | _ | 7 | Atr | _ | _ | +
-| 9 | التَبادُلِ_AltabAduli | تَبادُل_tabAdul | N | N | case=2<nowiki>|</nowiki>def=D | 8 | Atr | _ | _ | +
-| 10 | التِجارِيِّ_AltijAriy~i | تِجارِيّ_tijAriy~ | A | A | case=2<nowiki>|</nowiki>def=D | 9 | Atr | _ | _ | +
-| 11 | إِلَى_<ilaY | إِلَى_<ilaY | P | P | _ | 7 | AuxP | _ | _ | +
-| 12 | 500_500 | 500_500 | Q | Q | _ | 11 | Atr | _ | _ | +
-| 13 | مِلْيُونِ_miloyuwni | مِلْيُون_miloyuwn | N | N | case=2<nowiki>|</nowiki>def=R | 12 | Atr | _ | _ | +
-| 14 | دُولارٍ_duwlArK | دُولار_duwlAr | N | N | case=2<nowiki>|</nowiki>def=I | 13 | Atr | _ | _ | +
- +
-The first sentence of the CoNLL 2007 training data: +
- +
-| 1 | تَعْدادُ | تَعْداد_1 | N | N- | Case=1<nowiki>|</nowiki>Defin=R | 7 | Sb | _ | _ | +
-| 2 | سُكّانِ | ساكِن_1 | N | N- | Case=2<nowiki>|</nowiki>Defin=R | 1 | Atr | _ | _ | +
-| 3 | 22 | [DEFAULT] | Q | Q- | _ | 2 | Atr | _ | _ | +
-| 4 | دَوْلَةً | دَوْلَة_1 | N | N- | Gender=F<nowiki>|</nowiki>Number=S<nowiki>|</nowiki>Case=4<nowiki>|</nowiki>Defin=I | 3 | Atr | _ | _ | +
-| 5 | عَرَبِيَّةً | عَرَبِيّ_1 | A | A- | Gender=F<nowiki>|</nowiki>Number=S<nowiki>|</nowiki>Case=4<nowiki>|</nowiki>Defin=I | 4 | Atr | _ | _ | +
-| 6 | سَ | سَ_FUT | F | F- | _ | 7 | AuxM | _ | _ | +
-| 7 | يَرْتَفِعُ | اِرْتَفَع_1 | V | VI | Mood=I<nowiki>|</nowiki>Voice=A<nowiki>|</nowiki>Person=3<nowiki>|</nowiki>Gender=M<nowiki>|</nowiki>Number=S | 0 | Pred | _ | _ | +
-| 8 | إِلَى | إِلَى_1 | P | P- | _ | 7 | AuxP | _ | _ | +
-| 9 | 654 | [DEFAULT] | Q | Q- | _ | 8 | Adv | _ | _ | +
-| 10 | مِلْيُونَ | مِلْيُون_1 | N | N- | Case=4<nowiki>|</nowiki>Defin=R | 9 | Atr | _ | _ | +
-| 11 | نَسَمَةٍ | نَسَمَة_1 | N | N- | Gender=F<nowiki>|</nowiki>Number=S<nowiki>|</nowiki>Case=2<nowiki>|</nowiki>Defin=I | 10 | Atr | _ | _ | +
-| 12 | فِي | فِي_1 | P | P- | _ | 7 | AuxP | _ | _ | +
-| 13 | مُنْتَصَفِ | مُنْتَصَف_1 | N | N- | Case=2<nowiki>|</nowiki>Defin=R | 12 | Adv | _ | _ | +
-| 14 | القَرْنِ | قَرْن_1 | N | N- | Case=2<nowiki>|</nowiki>Defin=D | 13 | Atr | _ | _ | +
- +
-The first sentence of the CoNLL 2007 test data: +
- +
-| 1 | مُقاوَمَةُ | مُقاوَمَة_1 | N | N- | Gender=F<nowiki>|</nowiki>Number=S<nowiki>|</nowiki>Case=1<nowiki>|</nowiki>Defin=R | 0 | ExD | _ | _ | +
-| 2 | زَواجِ | زَواج_1 | N | N- | Case=2<nowiki>|</nowiki>Defin=R | 1 | Atr | _ | _ | +
-| 3 | الطُلّابِ | طالِب_1 | N | N- | Case=2<nowiki>|</nowiki>Defin=D | 2 | Atr | _ | _ | +
-| 4 | العُرْفِيِّ | عُرْفِيّ_1 | A | A- | Case=2<nowiki>|</nowiki>Defin=D | 2 | Atr | _ | _ | +
- +
-==== Parsing ==== +
- +
-Nonprojectivities in PADT are rare. Only 431 of the 116,793 tokens in the CoNLL 2007 version are attached nonprojectively (0.37%). +
- +
-The results of the CoNLL 2006 shared task are [[http://ilk.uvt.nl/conll/results.html|available online]]. They have been published in [[http://aclweb.org/anthology-new/W/W06/W06-2920.pdf|(Buchholz and Marsi, 2006)]]. The evaluation procedure was non-standard because it excluded punctuation tokens. These are the best results for Arabic: +
- +
-| Parser (Authors) | LAS | UAS | +
-| MST (McDonald et al.) | 66.91 | 79.34 | +
-| Basis (O'Neil) | 66.71 | 78.54 | +
-| Malt (Nivre et al.) | 66.71 | 77.52 | +
-| Edinburgh (Riedel et al.) | 66.65 | 78.62 | +
- +
-The results of the CoNLL 2007 shared task are [[http://nextens.uvt.nl/depparse-wiki/AllScores|available online]]. They have been published in [[http://aclweb.org/anthology-new/D/D07/D07-1096.pdf|(Nivre et al., 2007)]]. The evaluation procedure was changed to include punctuation tokens. These are the best results for Arabic: +
- +
-| Parser (Authors) | LAS | UAS | +
-| Malt (Nilsson et al.) | 76.52 | 85.81 | +
-| Nakagawa | 75.08 | 86.09 | +
-| Malt (Hall et al.) | 74.75 | 84.21 | +
-| Sagae | 74.71 | 84.04 | +
-| Chen | 74.65 | 83.49 | +
-| Titov et al. | 74.12 | 83.18 |+
  
 +Martin

[ Back to the navigation ] [ Back to the content ]